INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Heterotrimeric GTP-binding proteins (G proteins) of the G_q family activate the phosphatidylinositol-bisphosphate-inositol-trisphosphate-calcium pathway in a pertussis toxin-insensitive manner. Among G_q family members, G_q and G₁₁ share 88% homology and are expressed ubiquitously (2-4). G₁₄ is 81% identical to G_q and is restricted in its tissue distribution mainly to spleen, lung, kidney, and testes (5). G₁₅ and its human counterpart, G₁₆, show only 58 and 57% amino acid identity, respectively, to mouse G_q and are restricted to a subset of hematopoietic cells (5, 6). Upon activation, all G proteins of the G_q family activate  isomers of phospholipase C.

The coexistence of closely related G proteins in the same cell suggests different functions or different interactions between individual G proteins and either receptors or phospholipases. However, in reconstitution systems and transfected cells, G_q and G₁₁ have been shown to couple indiscriminately to a wide range of receptors, e.g. muscarinic m1 receptor (m1-R)¹ (7, 8) and m3-R (9), thyrotropin-releasing hormone (10, 11), parathyroid hormone (12), gastrin-releasing peptide (GRP) and vasopressin (13), gonadotrophin-releasing hormone (14), angiotensin II and bradykinin (15), histamine (16), and ₁-adrenergic receptors (17). G_15/16, despite restricted distribution, are capable of coupling many serpentine receptors tested, including those natively coupled to G_s and G_i (for review, see Ref. 18). Similarly, an attempt to assign different roles to G_q and G₁₁ using agonist-induced down-regulation failed to distinguish between the two G proteins (19-21).

This lack of discrimination may reflect real interchangeability in vivo among the G proteins of the G_q family or merely be a result of the assays used. To address this question, we have used the Xenopus oocytes system to study receptor-G protein specificity in vivo. Coexpression of thyrotropin-releasing hormone receptors with either mouse G_q or G₁₁ demonstrated different modulation of the response to thyrotropin-releasing hormone by each one of these G proteins (22). Applying the antisense approach, we demonstrated for the first time different coupling preferences of neuromedin B receptor for G_q and G₁₁ (1).

Although some information is available for other G_q family members, little is known of G₁₄-receptor coupling specificity. Wu et al. (17, 23) demonstrated in transfected COS-7 cells that G₁₄ mediates responses evoked by the ₁-adrenergic and the interleukin-8 receptors. Kuhn et al. (16) showed that histamine receptors couple to all G_q family members.

In this study, we have used the Xenopus oocyte system to investigate the coupling of native or expressed receptors to G_14. Xenopus oocytes express an endogenous protease receptor that is activated by trypsin (24). The response, typical for the activation of the phosphatidylinositol-specific phospholipase C signal transduction pathway, is manifested as Ca²⁺-dependent Cl current. Trypsin, along with several other proteases, stimulates fertilization-like responses in starfish eggs (25). In Xenopus oocytes, there is an elevation of G proteins of the G_q family during maturation and early development (26). These data suggest that protease receptors coupling via G proteins of the G_q family may be of importance in fertilization and early development. Using the antisense approach, we show here that G_q and G₁₄ are the main mediators of the response elicited by trypsin, in contrast to the response evoked by the activation of GRP-Rs, which utilize G_q and G₁₁, and the response to activation of m1-Rs, which utilize G_q and G_o.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Cloning of Xenopus G₁₄-- A cDNA library of Xenopus oocytes (constructed in gt10) was screened at low stringency with a mouse G_q probe, as described previously (1). Positive plaques were isolated and sequenced with universal and gene-specific primers using fMOL^TM sequencing kit (Promega), according to the manufacturer's instructions. A partial cDNA clone that displayed high homology to mouse G₁₄ was used as a probe in a second round of screening at high stringency: overnight hybridization at 37 °C in NT hybridization buffer (27) followed by a 4 × 10-min wash at room temperature in 1 × SSC (0.15 M NaCl and 0.015 M sodium citrate), 0.1% SDS and 20 min wash at 42 °C in 0.1 × SSC, 0.1% SDS. Positive clones were isolated and either subcloned into the EcoRI site of pGEM-4 plasmid (Promega) or PCR-amplified with pfu DNA polymerase (Stratagene) and sequenced on both strands with universal and gene-specific primers, as described above.

Antisense Oligonucleotides Sequence-- antiG_q: 5'-ATTCTCAAAAGAGGCGACC-3'; antiG₁₁: 5'-CTGTTCAAAGGTACATACT-3'; antiG₁₄: 5'-GTTTCCTTTCAAGACTGGAT-3'; antiG_o: 5'-GCGCTCAGTCTGCAGCCCAT-3'.

Handling of Oocytes-- Oocytes were excised from Xenopus females (South African Xenopus Facility, South Africa), defolliculated with collagenase, and kept in NDE solution at 20 °C, essentially as described previously (28). Stage V oocytes were injected with the desired RNAs (0.5-2 ng/oocyte) and/or phosphorothioate DNA antisense oligonucleotides (S-oligos, 50 ng/oocyte). The injections resulted in deterioration of some oocytes, particularly those injected with antiG₁₁. Viable oocytes were used in three assays: functional assay and Northern and Western analyses.

Functional Assay in Oocytes-- Measurements of electrophysiological responses to agonists were performed essentially as described previously (28). Briefly, individual oocytes were voltage-clamped at 70 mV. Agonists (trypsin, GRP, or acetylcholine (ACh)) were added directly to the bath, and membrane currents were continuously recorded. The mean amplitude of the control group in each experiment was set as 100%, and the effect of each treatment was calculated as the percent of that control response. N denotes the number of different donors, and n denotes the number of oocytes tested. Results are represented as means ±S.E. Paired Student's t test was used to estimate the statistical significance of the data.

Northern Analysis-- Groups of 80 oocytes were homogenized in 8 ml of cold guanidynium isothiocynate buffer, centrifuged for 5 min at 3000 × g, 4 °C, and the supernatant was loaded onto 4 ml of CsCl (5.7 M) cushion for standard total RNA isolation (27). The precipitated RNA was dissolved in 40 µl of H₂0. 6 µg of total RNA were loaded and resolved on RNA gel (agarose/formaldehyde) and blotted onto a nitrocellulose filter. Filters were air-dried and then baked at 80 °C in a vacuum for 1 h. Full-length clones of Xenopus G_q, G₁₁, and G₁₄ and partial clone of rat glycerol aldehyde phosphate dehydrogenase (GAPDH) were randomly labeled with ³²P (Random Primer labeling kit, Life Technologies, Inc.) and hybridized to the filters at high stringency (see cloning section above). Filters were autoradiographed for 3-7 days on XAR 5 Kodak films. Quantitation of the relevant bands was done by densitometry and normalization with the 28S and the 18S bands on the RNA gels and the 1.2-kilobase GAPDH bands on the autoradiograms.

Western Analysis-- Groups of 60 oocytes were homogenized in 1 ml of cold hypotonic membrane buffer (5 mM Hepes, pH 7.4, 5 mM NaCl, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 100 µg/ml soybean trypsin inhibitor) and centrifuged twice in a microcentrifuge for 5 s at 5,000 rpm to pellet the yolk. The supernatant was centrifuged for 20 min at 14,000 rpm. The pelleted membranes were resuspended in 60 µl of hypotonic membrane buffer and kept at 70 °C. Before use, loading buffer was added (1:1 v/v), and samples were heated to 90 °C for 5 min. 20 µl (equivalent to membranes of 10 oocytes) were resolved on 10% SDS-polyacrylamide gels and electrotransferred onto nitrocellulose filters. Filters were blocked for 3 h in blocking solution (TBS-T, Tris-buffered saline + 0.05% Tween 20, with 5% nonfat milk powder) and then exposed to the first antibody (1:500 dilution in blocking solution) overnight, with gentle shaking at 4 °C. Filters were then washed 5 times in washing solution (TBS-T) and exposed to the second antibody (1:1500 dilution in blocking solution of goat anti-rabbit IgG conjugated to horseradish peroxidase). Enhanced chemiluminescence system (ECL kit, Amersham Pharmacia Biotech) was used for developing the blots.

Materials-- Collagenase (Type 1A), trypsin (L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated), Hepes, aprotinin, leupeptin, soybean trypsin inhibitor, goat anti-rabbit IgG, and ACh were purchased from Sigma; GRP_14-27 from Bachem, S-oligos from Oligos Etc.; primers from Genemed Biotechnologies; and radiodeoxynucleotides from NEN Life Science Products. All molecular biology reagents were of molecular biology grade. All other chemicals were of analytical grade. WO82, B825, and Z808 antibodies were a gift of Dr. P. Sternweis, and QL antibody was a gift of Dr. T. Jones.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Cloning of Xenopus Oocyte G₁₄-- Initial screening of a Xenopus oocyte cDNA library with mouse G_q probe at low stringency resulted in the cloning of the amphibian counterparts of G_q, G₁₁, and an additional partial clone that most resembled mouse G₁₄ (see "Experimental Procedures"). This partial clone was used as a probe in a second round of screening at high stringency to isolate a full-length clone of Xenopus G₁₄. The cloned Xenopus G₁₄ cDNA (accession number AF059182) had an open reading frame of 1065 bases, predicting a 354-amino acid protein. The cDNA sequence was 78% identical to mouse G₁₄ (GenBank^TM accession number M80631). The predicted protein was 90% identical to its mammalian counterpart, with most differing residues representing conservative substitutions, except for a cysteine missing at position 4 (Fig. 1). The consensus sequences for the GTP binding site and the carboxyl terminus were identical to the mammalian protein. Northern analysis of the native RNA with the cloned G₁₄ cDNA as a probe detected a major 3.8-kilobase band and an additional 3.1-kilobase band (Fig. 2).

View larger version (81K): [in this window] [in a new window]   Fig. 1.   Alignment of the predicted amino acid sequences of Xenopus oocyte (upper row) and mouse (lower row) G14.

View larger version (34K): [in this window] [in a new window]   Fig. 2.   Specificity of antisense oligonucleotides in degrading cognate G protein mRNAs. Groups of oocytes were injected with each of the S-oligos (50 ng/oocyte). After 48 h, RNA was isolated, resolved on an RNA gel, and blotted onto nitrocellulose filters, as described under "Experimental Procedures." Filters were hybridized with 32P-labeled probes of Xenopus Gq (A), G11 (B), G14 (C), and mouse GAPDH (D) at high stringency. a-c, oocytes injected with antiGq, antiG11, and antiG14, respectively. d, control (uninjected) oocytes. In most of the experiments, lower molecular mass degradation products were observed for the targeted RNAs. E, densitometric analysis of relevant G protein mRNAs bands in three independent experiments after normalization with GAPDH bands. Percent remaining mRNAs: Gq (open bars); G11 (gray bars); G14 (black bars).

Antisense-induced RNA Degradation-- To test the antisense S-oligos selectivity for the corresponding mRNAs, we injected S-oligos into oocytes (50 ng/oocyte) and isolated RNA 48 h after injection. Northern analysis of three independent experiments after normalization with internal GAPDH transcript levels (see "Experimental Procedures") showed that all the three S-oligos were selective in promoting the degradation of their target mRNAs. Although antiG₁₁ and antiG₁₄ exhibited absolute selectivity, antiG_q was less effective and caused some degradation of G₁₄ (Fig. 2). 48 h post-S-oligos injection, antiG₁₄ degraded G₁₄ RNA by 86% and had no effect on G_q and G₁₁ RNA levels. antiG₁₁ completely abolished its cognate RNA and had no effect on the other two RNAs. antiG_q caused degradation of both G_q and G₁₄ transcripts by 62 and 42%, respectively, and had no effect on G₁₁ mRNA. The kinetics of RNA degradation show that most of the RNA is already degraded 3 h after antisense injection. Although a 3-h treatment with S-oligos did not always result in a complete disappearance of the G₁₁ or G₁₄ mRNA bands, low molecular mass products indicated substantial partial degradation (Fig. 3A). In a representative experiment designed to test early effects of antisense S-oligos (3 h), densitometry of the autoradiograms after normalization with endogenous GAPDH resulted in the reduction of G_q by 67%, G₁₁ by 89%, and G₁₄ by 77%. 24 h after the injection of S-oligos, the G₁₁ and G₁₄ mRNAs were fully degraded. Since some reports demonstrated recovery of the response 4-7 days post-antisense oligonucleotides injections (29, 30), we checked for possible reappearance of RNAs encoding these G proteins within the time frame of our experiments. Oocytes were injected with antisense oligonucleotide and subjected to RNA isolation 1, 2, 3, and 4 days post-injection and then to Northern analysis with the respective probes. No G₁₄-encoding RNA recovery was detected during this period (Fig. 3B). Similar results were obtained for the other two mRNAs (not shown).

View larger version (48K): [in this window] [in a new window]   Fig. 3.   Kinetics of mRNA degradation after antisense injection. 3-96 h post-S-oligos injection (50 ng/oocyte), RNA was isolated, resolved, blotted, and hybridized with 32P-labeled probes of Gq (q), G11 (11), and G14 (14), as described under "Experimental Procedures." +, antisense-injected oocytes; , control (uninjected) oocytes. A, 3 h post-injection of antiGq, antiG11, or antiG14; B, 1-4 days post-injection of antiG14. GAPDH Northern analysis for normalization was performed on the same filters after the decay of G protein probes.

Antisense-induced Depletion of Oocyte G Proteins-- To assess the effect of S-oligos-induced mRNA depletion at the protein level, we used Western analysis with antisera developed against the mammalian G proteins of the G_q family. Antibodies that specifically discriminate between mouse G_q and G₁₁ (WO82, B825; Refs. 15 and 31) interacted with a large number of proteins in the 30-50-kDa range and could not, therefore, detect amphibian G proteins unambiguously. Antibodies raised against the common carboxyl terminus of the G_q family (Z808 or QL, see Ref. 32) were more specific and, despite their interaction with a number of other proteins, we detected a faint band migrating at a position corresponding to approximately 42 kDa (Fig. 4). These data confirm the findings of Gallo et al., (26), showing only very small amounts of proteins of the G_q family in oocytes. The small amount of native G proteins made the quantitation of turnover rates difficult. To circumvent this problem, we overexpressed Xenopus G_q, G₁₁, or G₁₄ in oocytes by injecting (1 ng/oocyte) in vitro transcripts encoding their open reading frames. 48 h later, an increase in the intensity of the 42-kDa band was observed. The results indicated that G₁₁ and G_q were overexpressed, albeit to a different extent. G₁₄ did not seem to overexpress; however, injection of antiG₁₄ markedly decreased the signal. This could be interpreted either as the oocyte natively expressing mainly G₁₄ or that the expression of G₁₄ repressed the expression of other G proteins of the family. We have used oocytes injected with the transcripts of each of the three G_q family members to estimate their turnover rates. 24 h post-injection of the cognate S-oligos, membranes were subjected to Western analysis with QL and Z808 antisera. S-oligos treatments caused substantial degradation of each protein (Fig. 4A). Densitometric analysis of the overexpressed 42-kDa bands showed that G_q, G₁₁, and G₁₄ were degraded by 75, 50, and 50%, respectively (Fig. 4B). Since RNA degradation was very rapid, these numbers approximate the degradation rates of the overexpressed proteins. Mitchell et al. (33) found that the turnover of G_q/11 in CHO cells was best described by monoexponential curve with t1/2 = 18 h. Our data indicate similar G protein turnover rates in oocytes. In control oocytes, 24 h post-injection of antiG_q, the endogenous 42-kDa proteins were degraded by 30%, suggesting that native G proteins of the G_q family include approximately 40% G_q. These results taken together demonstrate that antisense S-oligos treatment caused rapid and significant depletion of oocyte G proteins.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   The effect of antisense oligonucleotides on immunoreactive proteins in oocytes overexpressing Gq, G11, and G14. Groups of oocytes were injected with each specific in vitro transcript (1 ng/oocyte) and 48 h later with the cognate S-oligo. 1 day later, oocyte membranes were prepared (see "Experimental Procedures"), resolved on SDS-polyacrylamide gels, and immunoblotted with QL antibody. A, immunoblot; B, densitometric analysis of the data in A. Br, membranes prepared from mouse brain; C, control (no RNA injected) oocytes, + or  antiGq; G14, oocytes injected with G14 in vitro transcript, + or  antiG14; G11, oocytes injected with G11 in vitro transcript, + or  antiG11; Gq, oocytes injected with Gq in vitro transcript, + or  antiGq. Similar results were obtained using the Z808 antibody.

The Effects of Antisenses on the Endogenous Responses to Trypsin-- To study the involvement of G_q, G₁₁, and G₁₄ in the response evoked by trypsin, oocytes were injected with each of the S-oligos (50 ng/oocyte) and challenged with 0.1-5 µg/ml (1-50 benzoyl-L-arginine ethyl ester units/ml) trypsin at different time intervals after the injections. 24 h after antiG_q injection, the response was reduced by 69 ± 4% (n = 57, N = 7, Fig. 5). antiG₁₄ reduced the response by 68 ± 7% (n = 58, N = 7). antiG₁₁ decreased oocytes viability, but the response in the surviving cells was not affected. Injection of only half of the amount of antiG₁₁ (25 ng/oocyte) did not compromise oocyte viability and had no effect on the response (81 ± 15% of control, n = 30, N = 3, not significant). In oocytes injected with a mixture of antiG_q and antiG₁₄, the response to trypsin was reduced to 7 ± 3% that of control (n = 57, N = 7). Despite the relatively high amounts of S-oligos injected in the mixture (50 ng/oocyte, each), oocyte deterioration was very limited and did not exceed that of oocytes injected with each one of the S-oligos alone. Similarly, membrane potentials and holding currents did not indicate any nonspecific functional damage due to the co-injection of the two S-oligos. These results suggest that G_q and G₁₄, but not G₁₁, are the main mediators of the endogenous response to trypsin in Xenopus oocytes.

View larger version (13K): [in this window] [in a new window]   Fig. 5.   The effect of antisense oligonucleotides on responses to trypsin. Oocytes were injected with antiGq (q), antiG14 (14), antiG11 (11) (50 ng/oocyte), or a mixture of antiGq/antiG14 (q/14) (50 ng/oocyte each antisense). 24 h later, oocytes were tested for responses to 1 µg/ml trypsin. Upper panel, tracings of representative responses; lower panel, average responses in seven independent experiments (n = 58). Data were calculated as percent of control (C) responses in each experiment. AS, antisense-injected.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Time and dose dependence of the effect of antisense oligonucleotides. Oocytes were injected with antiGq (q), antiG14 (14), or a mixture of the two (q/14), as described in legend to Fig. 5, and tested 24 and 48 h later for their response to either 1 or 5 µg/ml trypsin. AS, antisense-injected; Time, time after antisense injection; [Tryp], trypsin concentration (µg/ml). The mean control response to 1 µg/ml trypsin after 24 h was 1388 ± 161 nA (n = 14) and after 48 h, 1274 ± 225 nA (n = 10). Mean control response to 5 µg/ml after 48 h was 2857 ± 410 nA (n = 15). A representative experiment is shown.

View larger version (28K): [in this window] [in a new window]   Fig. 7.   The effect of antisense oligonucleotides on GRP-R- and m1-R-induced responses. Oocytes were injected with in vitro transcripts encoding the GRP-R and m1-R and, 18 h later, with antiGq (q), antiG11 (11), antiG14 (14), or antiGo (o) (50 ng/oocyte). 24 h after the injection of S-oligos, oocytes were tested for responses to 1 µM GRP or 10 µM ACh. Responses are presented as % of control responses in each experiment. *, p < 0.05.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

All G proteins of the G_q family have been shown to couple to GPCRs in the pertussis-toxin-insensitive phosphatidylinositol-specific phospholipase C signal transduction pathway. With the exception of our previous report in oocytes (1), transfection and reconstitution experiments failed to functionally discriminate between G_q and G₁₁. However, sequence and distribution differences (e.g. Ref. 4) between the two G proteins as well as among all the family members in general suggest individual physiological roles in regulating signaling in target tissues and organs. The antisense approach has the advantage of relying on the natural milieu of each component of the signal transduction pathway. By applying this approach in Xenopus oocytes, we have previously demonstrated functional in vivo preferences of neuromedin B receptor coupling via G_q (1).

To broaden the scope of the study, we have cloned Xenopus G₁₄ in order to develop an antisense oligonucleotide that could be used to deplete cells of G₁₄ protein. Similarly to the previously cloned Xenopus G_q and G₁₁, Xenopus G₁₄ showed a high degree of homology to its mammalian counterpart, demonstrating the conservation of these protein sequences during vertebrate evolution. Northern analysis in oocytes demonstrated the intrafamily selectivity of the designed antisenses (antiG_q being less selective, degrading mainly G_q but also G₁₄) and the rapid elimination of the cognate mRNAs after antisense injection. Thus, proteins turnover rates were the limiting factor in the depletion process. The turnover rate of G_q/₁₁ has been shown in Chinese hamster ovary cells to fit a first-order reaction with a t1/2 of 18 h (33). Our data in oocytes, obtained from both Western analysis and functional assays, exhibit similar kinetics of degradation, but distinguish between G_q and G₁₁. The turnover rate of G₁₄ was similar to that of G₁₁. Our functional assays were executed 24-80 h after antisense injection, a time window in which most of the G_q and more than 50% of G₁₁ and G₁₄ were degraded. Therefore, a 40% reduction in the m1 muscarinic response by either antiG_q or antiG₁₁ and the fact that this response was not further reduced after 48 and 80 h support the existence of an additional mediator for this receptor. Indeed, antiG_o caused a 20% reduction in the response to ACh 24 h after its injection. Mixtures of antiG_q/antiG₁₁ caused pronounced oocytes deterioration, preventing a more conclusive interpretation. We cannot exclude the possibility that even residual amounts of any G protein (undetectable by Western analysis) can mediate a full response, provided there is a large excess of G protein over the receptor and high affinity between the two proteins.

The antisense oligonucleotides designed against the different G proteins could have achieved their effects by interfering with the synthesis of different phospholipase C isoforms. This possibility, however, appears unlikely since there is no evidence for multiple phospholipase C species in oocytes.

Since there is little information about the G₁₄ ability to couple GPCRs, we tested a number of receptors (mouse m1, GRP and neuromedin B receptors) and the native trypsin response to investigate its specificity. Only the endogenous protease receptor exhibited sensitivity to G₁₄ depletion. Two antisense oligonucleotides, antiG_q and antiG₁₄, each diminished the responses to about 30%, and a mixture of the two reduced it to 7% that of control, whereas antiG₁₁ had no effect. These results strongly suggest that the response to trypsin is mediated exclusively by G_q and G₁₄. Although it appears that the contribution of these two G proteins to the trypsin response is similar, the limited degree of degradation of G₁₄ mRNA caused by anti G_q precludes a firm conclusion regarding this issue. Thus, G₁₄ mediates the response to protease in addition to its ability to couple ₁-adrenergic, interleukin-8 receptors (17, 23), and histamine receptors (16). Wu et al. (37) have shown by altering selected domains within the third intracellular loop of the _1B-adrenergic receptor, sequence-related specificity for either G_q/₁₁ or G₁₄. It would have been interesting to compare the sequence of the yet uncloned Xenopus protease receptor, which appears to couple to G_q and G₁₄, but not to G₁₁, with that of the _1B-adrenergic receptor.

The finding that antiG₁₁ did not affect the response confirms our previous report that in vivo G_q and G₁₁ have distinct receptor preferences. It is interesting to note that so far in reconstitution, transfection, and in vivo models, every GPCR tested (with the possible exception of neuromedin B receptor expressed in oocytes, Ref. 1) was able to couple to more than one G protein of the G_q family. Since different GPCRs appear to be able to discriminate among the G proteins, the interaction of a single receptor with more than a single G protein implies biological significance. In this respect, the recent report by Macrez-Lepretre et al. (38) points to discrimination between G_q and G₁₁ for specific effector molecule in the signaling pathway. The G proteins specificity for the receptors on the one hand and for the distal parts of the signal transduction pathway on the other may serve to fine tune receptor-mediated biological responses.